http://purl.org/dc 
  Dublin Core Metadata Initiative
 
 
Home Search Site Map What's New Feedback
 
  Questions and Answers / FAQ  
 
QUESTIONS AND ANSWERS
FAQ
Glossary
Usage Guidelines
   

FAQ

Last Updated 2000-03-28

What is Metadata?
What is the purpose of Dublin Core metadata?
Why is it called the "Dublin Core"?
What is "Simple Dublin Core"?
What is "Qualified Dublin Core"?

Who uses Dublin Core?
Is Dublin Core metadata intended to be used only for digital (electronic) resources?
What is the Warwick Framework?
How can I store Dublin Core Metadata?
How can I store Dublin Core Metadata in HTML?
What is the relationship between Dublin Core Metadata and RDF and XML?
What is the Dublin Core datamodel?

Can I add a new element to Dublin Core?
How do I store names in Dublin Core?

What is the maximum length for each field in Dublin Core?
Please explain the one to one (1:1) rule.
What search-engines support Dublin Core?
How do I participate in discussion about Dublin Core?
Workshops?
What is the relationship between Dublin Core and other Internet standards groups?
Is there metadata on the Dublin Core Metadata Initiative's web site pages?

What is Metadata?

The simplest useful definition of metadata is "structured data about data." This very general definition includes an almost limitless spectrum of possibilities ranging from human-generated textual description of a resource to machine-generated data that may be useful only to software applications..

The term metadata has been used only in the past 15 years, and has become particularly common with the popularity of the World Wide Web. But the underlying concepts have been in use for as long as collections of information have been organized. Library catalogs represent a well established variety of metadata that has served for decades as collection management and resource discovery tools.

What is the purpose of Dublin Core metadata?

Dublin Core metadata is specifically intended to support resource discovery. The elements represent a broad, interdisciplinary consensus about the core set of elements that are likely to be widely useful to support resource discovery.

Why is it called the "Dublin Core"?

The original workshop was held in Dublin, Ohio, hence the term "Dublin Core."

The DC directorate is located in Dublin, Ohio at the Online Computer Library Center, a nonprofit organization that serves the networking needs of libraries in the US and some 70 other countries. The Dublin Core Metadata Initiative itself, however, is very much international in scope.

What is Simple Dublin Core?

"Simple Dublin Core" is a term often used to describe Dublin Core metadata that uses no qualifiers (see "Qualified Dublin Core" below). That is, the elements are expressed as attribute-value pairs using just the 15 elements from the Dublin Core Metadata Element Set without any further information about encoding schemes, enumerated lists of values, or other processing clues.

The term "unqualified Dublin Core" is synonymous with "simple Dublin Core."

What is Qualified Dublin Core?

"Qualified Dublin Core" is a term applied to Dublin Core metadata that employs additional information to increase the specificity of the metadata by refining the meaning , by specifying encoding schemes or controlled vocabularies, or to indicate a metadata value is a compound, or structured value.

For example, a date may be further identified as a particular variety of date (date last modified, date published, etc.) and might be encoded according to a particular scheme that assures that it can be interpreted unambiguously. A subject term that is the value of the Subject element might be specified as having been selected from a particular controlled vocabulary such as the Dewey Decimal Classification.

Qualifiers allow applications to increase the specificity or precision of the metadata, but also introduces complexity that can impair interoperability. Thus, designers of metadata systems are well advised to be conservative about deploying qualifiers in cases where interoperability is one of the design objectives.

Dublin Core interoperability qualifiers are those which have been approved in the Dublin Core community and are a formal part of the registry of Dublin Core metadata. Designers should select qualifiers that come from this set to the extent that they meet the functional requirements for a given application.

It is expected that local or application-specific requirements may well require additional qualifiers, or even additional elements, that have not been approved by the Dublin Core community at large. Deployment of local qualifiers and extensions is an appropriate action, but designers should do so with the understanding that interoperability with other applications may suffer. In cases where such additional qualifiers are being deployed, it may be helpful to bring this to the attention of the Dublin Core Directorate in order to promote the wider use of such qualifiers.

Who uses Dublin Core?

Anyone can use Dublin core elements as a convenient basis for descriptive systems. Web pages are one of the most common types of resources to have Dublin Core descriptions, often using HTML meta tags. But bulk conversions of descriptive information stored in older catalog systems and databases probably made up the largest number of Dublin Core descriptions in the early days of the initiative. Dublin Core metadata is being used as the basis for descriptive systems by several specialised communities and agencies: community based interest groups currently exist for education, libraries, government, and research. Also see the listing of Dublin Core based projects.

Is Dublin Core metadata intended to be used only for digital (electronic) resources?

The Dublin Core Metadata Initiative began in 1995 to develop conventions for resource discovery on the World Wide Web. The focus of discussion was electronic resources. It was clear at the outset, however, that the semantics of resource discovery should be independent of the medium of the resource, and that there are obvious advantages for using the same semantic model across media. Thus, considerable attention has been invested in making the Dublin Core sufficiently flexible to represent resources (and relationships among resources) that are both digital and exist in traditional formats as well.

What is the Warwick Framework?

The Warwick Framework is a conceptual model for metadata that recognizes that different packages of metadata, about the same thing, might be useful for different purposes and created, according to different standards or by different agencies. These will coexist, and mechanisms for generating and storing metadata should make allowances for this requirement. XML namespaces are one example of a tool which underpins the Warwick Framework approach.

The Warwick Framework is not a system, but rather a set of design principles that have helped guide the development of the Dublin Core since the Second Dublin Core Metadata Workshop in Warwick, UK. Please read "The Warwick Metadata Workshop: A Framework for the Deployment of Resource Description" for more information.

How can I store Dublin Core Metadata?

The Dublin Core Metadata Element Set is a set of semantic definitions of 15 descriptive elements. This means that many different methods can be used to actually record or transfer Dublin Core metadata. Common methods include HTML, XML, RDF and relational databases.

Nevertheless, there are various limitations associated with the different methods that should be taken into account if planning a system.

For example, HTML is convenient since it keeps the metadata embedded together with a document content, but only supports a simple list of meta elements, with no grouping or other structure.

XML allows much more structured data to be recorded, and mechanisms such as XML namespaces can add value. However, there are few software applications available to use XML encoded data yet, and those that exist require considerable configuration or programming.

RDF, which was developed partly by people from the Dublin Core community, probably has the greatest expressive potential, but software to support RDF is currently experimental.

For best performance in searching and retrieval, it is probably best to load DC metadata into a database management system, but this then forces the detachment of the metadata from the resource which it describes, so system designers must be careful to implement procedures to ensure consistency and synchronisation.

How can I store Dublin Core Metadata in HTML?

Dublin Core metadata can be stored using the meta element in the head of HTML documents. An information "RFC" issued by the IETF Encoding Dublin Core Metadata in HTML defines the standard way to do this. A Note issued by DCMI describes one method for storing Qualified DC in HTML.

What is the relationship between Dublin Core Metadata and RDF and XML?

Dublin Core Metadata is about semantics... what we are trying to say about resources.

RDF is about structure... the conventions for encoding the assertions about a resource that uses Dublin Core Metadata semantics.

XML provides the syntax for encoding assertions in RDF, and hence in RDF-encoded Dublin Core Metadata.

Dublin Core Metadata and RDF are distinct specifications. Neither requires the other, but their co-evolution make them natural complements in the metadata architecture for the Web. The Dublin Core community and the RDF community have a number of members in common, and they have evolved together. The Dublin Core community, in fact, provided many (but not all) of the basic requirements that were used to design RDF. In turn, the development of RDF provided the Dublin Core community with a much more formal underlying data model which has helped the DC community find principled solutions (rather than ad hoc Band-Aids) for many of the detailed problems that arise in deployment.

What is the Dublin Core datamodel?

A datamodel is an abstract representation of the part of the world that a system is designed to service. It typically includes its scope - i.e. an indication of the range of objects of interest to the system; the properties of things which are in-scope, and some kind of specification of interactions between them.

The RDF datamodel is quite simple

  • "things have properties", where each property has a name and a value. The "atom" of RDF, therefore, has three components: the thing (resource, in web terminology), a property name, and a property value.
  • each resource may have several of these properties.
  • the value may itself be a resource.

Dublin Core builds on RDF by:

  • defining 15 named properties - the Dublin Core Metadata Element Set - which are considered to be widely useful for resource discovery.
  • specifying that a resource may have any number, including zero, of each of these properties (elements, in DC terminology).
  • defining the range of resource types that may have DC descriptions - given by the allowed values for the DCMES Type element.

which defines the basic Dublin Core datamodel.

Several of the 15 DC elements record relationships between the resource-of-interest and another resource which is in-scope for DC. In particular, Relation and source are used to indicate a connection with another resource of any type. Creator, Contributor and Publisher relate the present resource to a party who has some responsibility for it. Coverage relates the present resource to a place or to a time-period. The Value recorded for each of these elements, therefore, should normally be strictly an identifier of another resource, which could, in principle, have its own DCMES description. In this way, it is possible that a web of DC descriptions of resources will be constructed, with each resource at the centre of a fan of descriptive arcs, but connected via one of these relationships to other resources with their own fans. Of course, the "identifiers" will, in many cases, use an informal notation in plain-text.

However, many users have found it useful to add extra information to Dublin Core descriptions, in particular:

  • values are frequently drawn from a controlled vocabulary (e.g. a keyword list, such as Library of Congress Subject Headings, LCSH), or written using a special notation (e.g. the ISO8601 format for dates and times), or use a particular natural language (in the case of values written in text-strings), so it would make the information more useful if a client were informed of the source of this vocabulary, the definition of the notation, or the name of the language used.
  • an instance of an element may be used in a specialised way, with more restricted semantics than implied by the broad definition of the element (e.g. the Date might be the date a resource was modified)

To encompass these requirements, the qualified Dublin Core datamodel also includes:

  • value qualifiers - to store an identifier for the vocabulary, encoding or language of the value.
  • element qualifiers - to refine the meaning of the element. A DCMES element qualified with an element qualifier is, effectively, a "new" element or property, with a more specialised meaning than its parent element.

Note, however, that it is not possible to create a new Dublin Core element with a meaning that goes beyond the meaning of the meaning of elements in the DCMES. To extend a description beyond that supported by DCMES, other mechanisms must be used.

Can I add a new element to Dublin Core?

As its name implies, the Dublin Core is a starting place, not a final word. It is expected that local applications will add elements to meet local or community needs. Local extensions should be clearly distinguished from the Dublin Core through the use of a notational mechanism, such as using a prefix or namespace. Elements which are part of a local extension will not normally be useful to others that they have not been subject to the cross-community consensus building process that has brought the Dublin Core element set to its current state. Implementors are, however, encouraged to bring such elements to the attention of the community so that others might adopt them as judged necessary, leading possibly to eventual approval as additional Dublin Core elements.

The Australian Government Locator Service is an example of a major implementation of Dublin Core metadata which has added four local elements to Dublin Core in the recommended way.

How do I store names in Dublin Core?

The encoding of personal names is fraught with many difficulties in Dublin Core metadata, as in other types of metadata systems. Conventions differ from culture to culture. The recommendation for Dublin Core metadata is to encode family name first, which supports effective collation of names and is consistent with most naming conventions globally. Two Notes have been written about the difficulties of personal names among cultures. One was written by Diane Hillmann (Names in Dublin Core) and the other by Andrew Waugh (Representing People's names in Dublin Core).

What is the maximum length for each field in Dublin Core?

There are no limits to field length.

Please explain the one to one (1:1) rule.

The 1:1 rule expresses the notion that a discrete resource should have a discrete metadata record. Seems simple enough. Unfortunately, it isn't always so neat because resources are often not so discrete (Should each photograph in an article have its own record? What about collections of articles? Can the collection be thought of as a resource? What about multimedia objects?)

Some in the Dublin Core community now believe that the 1:1 principle should be revisited, and a working group will address these issues and recommend guidelines to promote consistent application.

If you are interested in joining this group (dc-one2one), either as an active participant or as a listener, please visit the Dublin Core working group area on the Mailbase Server (http://www.mailbase.ac.uk). The site provides directions for joining (and leaving) mailing groups. Names of all Dublin Core mailing lists begin with dc- and can be found under D in the list of groups.

Which search-engines support Dublin Core?

Several commercial and noncommercial search-engines will index meta elements with just a little configuration. A recent inquiry on the dc-general mailing list produced this list:

To get a good overview of what software is out there see Search Tools and Search Engine Watch.

The famous "all the web" search engines like AltaVista, Yahoo, HotBot, etc. tend to avoid using the information found in meta elements in their indexing. This is because, unless the pages are from guaranteed "trusted" servers, the meta information is commonly used by unscrupulous content-providers for spamming, to mislead the indexes into giving web-pages a misleading rating.

How do I participate in discussion about Dublin Core?

You can sign up for any of the DC mailing lists or working groups at the Mailbase server. Go to: http://www.mailbase.ac.uk and follow the instructions for joining a mailing list. The names of all Dublin Core mailing lists start with dc- (click on Discussion Lists and then on D).

Workshops?

Information will be posted on the Dublin Core Homepage (http://purl.org/dc) on how to nominate yourself for upcoming workshops. Please note that workshops are not introductory workshops, but are intended for those who are implementing Dublin Core and are interested and able to contribute time and effort in advancing the state of the art.

What is the relationship between Dublin Core and other Internet standards groups?

The Dublin Core Metadata Initiative (DCMI) is a consensus building organization that has relationships in many standards activities. A number of people in the DCMI are active in the W3C (DC is the prototype application that drove the development of the Resource Description Framework, or RDF in the W3C). Our own standardization activities take place in the IETF (RFC 2413 is reference description of the initial version of the Dublin Core), and there are currently formal DC standardization activities underway in CEN (the European information industry standardization forum) and in NISO (the North American information standardization organization).

Is there metadata on the Dublin Core Metadata Initiative's web site pages?

All of the DC pages have metadata associated with them. For practical management purposes, we've associated external metadata linked by the HTML LINK element.

When you view source on http://purl.org/dc/index.htm you'll find
<LINK REL="meta" HREF = "index.htm.rdf" />, take a look at http://purl.org/dc/index.htm.rdf and you'll find what you're looking for.

 
 
 

Dublin Core Contact
For questions or comments regarding the Dublin Core contact dc@oclc.org

     
       
Home | Search | Site Map | What's New | Feedback | About the Dublin Core | News and Publications | Documents | Questions and Answers | Projects | Tools | Working Groups | Workshop Series
 
           
© 2000 DCMI